Existing search engines typically display a list of search results associated with a search query as a list of relevant web pages. This list may include web pages with identical or similar content. For example, when a search query matches a particular section of a web page, a user is typically exposed to many copies of the same or similar information. Some existing approaches involve summarizing a single document, for example, by choosing particular sentences from the document, rather than by presenting the information as a whole from a set of documents. It would be desirable to develop a technique by which results of a search query can be grouped efficiently so that the duplicate content appearing to the user is minimized.
Further, existing approaches to refine a search associated with a search query typically involve entering a new search query. Some search engines include suggested topics in response to a search query. However, these suggestions are generally based on criteria, such as popularity or past search criteria. It would be desirable to develop a technique by which query refinements for a search query can be automatically generated from a search result set.
The current state of the art in Web advertising relies upon relatively simple advertisement placement paradigms. Current approaches to web advertising typically involve displaying sponsored link advertisements or banner advertisements for a given search term. Sponsored link advertisements are generally listed in an order determined by the search engine, typically by some combination of pay-per-click bid auction and relevancy factor. Currently, search engines enable advertisers to pay, usually by bidding for sponsored link placements along with non-sponsored search results or for keywords to which the advertisers want to match sponsored link advertisements. Refining an advertisement typically involves entering a new bid on the keyword query extension. Exisiting approaches may provide suggestions to advertisers on which additional keywords and query extensions to bid. Even with these suggestions, advertisers typically guess the keywords that users may use to refine their searches and then bid on those keywords. Furthermore, these suggestions are typically based on popularity, past searches or other criteria, so that bidders who bid on these keywords have a possibility of their advertisement being displayed.
It would be desirable to develop a technique by which advertisements can be automatically generated from the search results associated with a search query. In addition, it would be desirable to develop a technique by which the placement of advertisements is not directly related to the bidding on specific keywords or query extensions to those keywords, but can be determined based upon criteria entered by an advertiser.